Uptake of uranium in Atlantic salmon (Salmo salar) : aqueous exposure

(1)

Master’s Thesis 2020 30 ECTS

The Faculty of Environmental Sciences and Natural Resource Management

Uptake of uranium in Atlantic

salmon (Salmo salar) – aqueous

exposure

(2)

(3)

Uptake of uranium in Atlantic salmon (Salmo salar) – aqueous exposure

The Faculty of Environmental Sciences and Natural Resource Management (MINA) Norwegian University of Life Sciences (NMBU)

Aas, Norway

BIRGITH ØVERBY TERUM

Aas, Norway February 2020

(4)

(5)

Acknowledgements

This thesis represents an output of a five years master study in chemistry and biotechnology (civil engineering): Environmental chemistry in the Department of Chemistry, Biotechnology and Food science (KBM) at the Norwegian University of Life Science (NMBU)

My most profound gratitude goes to my supervisors, Hans-Christian Teien and Brit Salbu, and the Ph.D. candidate Shane Scheibener, who gave me the distinct privilege to be a part of their study of aqueous exposure of waterborne uranium in Atlatic salmon (Salmo salar). I have learnt so much during this time, and my interest in ecotoxicology has flourished. Thank you for your solid patience and for taking the time to provide thorough explanations to my questions.

Many thanks to the staff of the Environmental Chemistry Section of the Faculty of Environmental Sciences and Natural Resource Management (MINA).

Thanks to Hans-Christian Teien! For all the support, thorough explanations and motivational pep-talks during the process of writing this thesis. I am so grateful. I have learnt a lot from you, and I appreciate all the time you have invested in this thesis.

Thanks to Shane Scheibener! You have patiently answered all my questions, and friendly letting me take part in your study. I have learnt so much during the work with this thesis.

Good luck in your continued work with your Ph.D. You will do great!

Thanks to all of the staff at the Isotope Laboratory! Karl-Andreas, Estela and all the others, you have helped me when I felt stuck in the data analysis and made me feel very welcome in the office.

To my loving family and friends, thank you for all your support during my study and especially the last seven months. I would not have made it without you.

Birgith Øverby Terum, Msc.

(6)

Statement of declaration

The thesis is the final part of the civil engineer study in chemistry and biotechnology, specialization in Environmental Chemistry, at the Norwegian University of Life Sciences (NMBU), Aas, Norway. The thesis is a part of a larger Ph.D study conducted by Shane Scheibener. He had the full responsibility for the experiment and lead the work completely. I participated in the set-up of exposure tanks, dissection of the exposed fish tissues (gills, kidney, liver, stomach, brain, muscle and bone) two days in June and two days in August), and some day-to-day work like feeding, water exchanges, water sampling, onsite water fractionation and daily measurements of water quality. I participated in the pre-treatments of the tissue samples, i.e. freeze-drying, weighing and preparing for digestion. I took part in the analysis on ICP-MS of both water and tissue samples.

Permission to copy in whole or in part of this thesis should be addressed to:

The Faculty of Environmental Sciences and Natural Resource Management (MINA) Norwegian University of Life Sciences (NMBU)

NMBU - MINA

(7)

List of acronyms

Bq: Becquerel EtOH: ethanol

oC: degrees of Celsius Ca: calcium

CO2: carbon dioxide

CRM: certified reference material DO: dissolved oxygen

DOC: dissolved organic carbon DOM: dissolved organic material DU: depleted uranium

dw: dry weight F: fluoride fw: fresh weight g: gram

HCl: hydrochloric acid He: helium

HNO3: nitric acid

ICP-MS: inductive coupled plasma – mass spectrometry i.e.: id est (in other words)

K: potassium kDa: kilo Dalton L: litre

LMM: Low Molecular Mass LOD: level of detection LOQ: level of quantification Mg: magnesium

N: number Na: sodium NH3: ammonia NH4+: ammonium NO3-: nitrate

NORM: naturally occurring radioactive material

(8)

P: Phosphorus Pb: lead

pH: the negative log of the concentration of the hydrogen cation (pH = -log10[H⁺]) PO43-: phosphate

Pu: plutonium Ra: radium Rn: radon

RO-water: reverse osmosis water RSD: relative standard deviation S: sulphur

SO42-: sulphate Th: thorium

TOC: Total organic carbon U: uranium

µ: micro

UF: Ultrafiltration UP: ultrapure

US- EPA: United States Environmental Protection Agency vv: volume volume (concentration)

ww: wet weight

(9)

List of tables

Table 1.1: Overview of quantified BCFs at Kazakhstan, Tajikistan and Kyrgyzstan. ... 11

Table 3.1: Overview of the reference material used when measuring the water samples with certified concentrations and determined concentrations of house standard and CRM. Values presented in average ± std.dev. Calculations of RSD and bias in appendix 2 (eq. 7 and 4). ... 21 Table 3.2: Certified concentrations and determined concentrations of the reference material, in addition deviation (bias, %) in certified reference material used when analysing fish samples. Values presented in average ± std.dev. Calculations of RSD and bias in appendix 2 (eq. 7 and 4). ... 22 Table 3. 3 The average temperature, pH, conductivity, DO-, ammonium, CO2-levels and ion composition in the exposure-waters. Concentrations were quantified using <0.45 µm-filtered water samples. Values are presented in average ± std.dev (mg/L). ... 23 Table 3.4: Concentrations of the paired fractions of uranium through the exposure. The percentage of fraction of the total concentration are listed in the parenthesis. LOD = 0.0003.

... 24 Table 3.5: Average concentrations in different tissues after 28 days of exposure. ... 29

Table 3.6: Correlation between possible uptake-routes (concentration) of U (gills, skin, stomach) and accumulation (concentration) in internal organs (kidney, liver and muscle).

Concentrations were plotted in µg/g tissue. The U concentration were compared between two organs in the same fish. Correlations between U accumulation in internal organs (liver- kidney, liver-muscle, kidney-muscle) were quantified, as well as correlation between uptake- routes (gill-skin and gill-stomach). ... 40

(12)

List of figures

Figure 1.1: The graphs indicate the fraction of total UO2 bound to DOC at different

concentration affected by pH (Goulet et al., 2012). ... 4 Figure 1.2: Speciation of LMM-U linked to pH. Uranium concentration in A) is 1nmol/L an B) is 110 nmol/L (Goulet et al., 2012). ... 5 Figure 1.3: Overview of fractions (Salbu et al., 2015) ... 6

Figure 2.1: Fish (n=50) were transferred from the breeding-tank to a separate exposure- tank.

The tanks were connected by tubes to maintain a continuous waterflow and a stable level of water in the exposure tank. ... 15 Figure 2.2: The left picture, length was quantified of fish. Right picture shows the fish before the dissection of internal organs (Pictures: B. Terum, 11.06.19). ... 18 Figure 3.1: Pie-diagram of the fractional- distribution at day 2, 17 and 28. 1 - particulate U (blue), 2 - Colloidal U (orange), 3 – LMM-U (grey). Particulate U (n=4), colloidal U (n=2), LMM-U (n=2). ... 26 Figure 3.2: The average concentrations of uranium per day, in both treatments, were graphed.

It was an even amount of time-points during the exposure. The “change of water”-days were marked as black squares. Error-bars with the individual standard deviation for each point were plotted. Each time-point was based on the average ± standard deviation each day. ... 27 Figure 3. 3 Overview of the accumulation of U in gills during the exposure. The fish were exposed to an average of 50.1 ± 21.0 µg/L U (based upon <0.45 µm-filtered water samples (n=48). The graph “First order kinetics” image exponential rise to a maximum, predicted by SigmaPlot. ... 30 Figure 3. 4 Overview of the U accumulation in liver during the exposure. The fish were exposed to an average of 50.1 ± 21.0 µg/L U (based upon <0.45 µm-filtered water samples (n=48). The graph “First order kinetics” image exponential rise to a maximum, predicted by SigmaPlot. ... 32 Figure 3.5: Overview of the U accumulation in kidney during the exposure. The fish were exposed to an average of 50.1 ± 21.0 µg/L U (based upon <0.45 µm-filtered water samples (n=48). ... 33 Figure 3.6: Overview of the U accumulation in skin during the exposure. The fish were exposed to an average of 50.1 ± 21.0 µg/L U (based upon <0.45 µm-filtered water samples (n=48). ... 34 Figure 3.7: Overview of the U accumulation in stomach (w/content) during the exposure. The fish were exposed to an average of 24.9 ± 11.9 µg/L U (based upon the average particulate U concentration in the water (n=6). The graph “First order kinetics” image exponential rise to a maximum, predicted by SigmaPlot. ... 36

(13)

Figure 3.8: Overview of the U accumulation in muscle during the exposure. The fish were exposed to an average of 50.1 ± 21.0 µg/L U (based upon <0.45 µm-filtered water samples (n=48). ... 37

(14)

List of appendices

Appendix 1: U distribution in fish species (Kazakhstan, Tajikistan and Kyrgyzstan) ... 51

Appendix 2: Details of analysis and calculations examples... 52

Appendix 3: Uptake rate ... 55

Appendix 4: Calculation of correlation factor ... 56

(15)

Abstract

The MSc. thesis assess the ecotoxic effects of uranium in freshwater ecosystems, by analysing the possible uptake-routes of U in fish using juvenile Atlantic salmon (Salmo salar) as a model organism. Bioconcentration factor (BCF), uptake-rates and organ distribution of accumulation were used to determine the potential U uptake.

Objectives: (1) There is an uptake of uranium in the fish directly from water. (2) The concentration of uranium in gills is higher than in stomach at steady state in aqueous

exposure. (3) Uptake-rates in gills and skin, compared to muscle, kidney and liver are higher;

Atlantic salmon (Salmo salar) (n=50) were exposed to waterborne uranium in moderate hard US-EPA water for 28 days. A stable pH (6.7) and water conditions were maintained in the exposure to keep the speciation of U constant. Water samples were collected every day and water fractionation-samples were collected three times during the exposure (day 2, 17 and 28). Fish were sampled five times during exposure using the EMERGE protocol (Rosseland et al., 2001). The bioaccumulation of U in the different tissues was quantified. Elemental

concentrations were quantified using Agilent Technologies 8800- QQQ ICP-MS. The precision and accuracy were considered based upon the analysis of certified reference materials, estimated relative standard deviation (RSD), and quantification and detection limits.

The general water quality was not significantly different between the control- and 50 µg U/L- treatment. The average concentration in the nominal 50µg/L experiment was 55 ± 22 µg U/L (n=32) and 0.027 ± 0.038 µg U/L (n=38) in the control exposure waters. The low molecular mass (LMM) of uranium was present as 19 % of total U concentration. The assumed

bioavailable (U-cations) concentration was on average 1.13 ± 1.74 µg U/L of total U concentration. U spikes, water exchanges and rapid sorption of bioavailable U to dissolved organic matter (DOM) in the water did likely cause a dynamic change in fractionation of U.

The accumulation of uranium was significant in all of the organs analysed after 28 days of exposure. It was no significant U uptake in the organs from the control-exposure. The highest U concentration and uptake-rate were quantified in the gills (5.9 ± 0.9 µg/g tissue ww and 0.22-0.24 µg U/g tissue/day). The U concentrations were as follows: gills>stomach w/content

>skin>kidney>muscle>liver. The uptake-rates were significantly higher in the organs with

(16)

direct contact with the contaminated water, compared to the internal organs analysed (muscle, kidney and liver). The apparent BCF in liver was 0.52 L/mg (<0.45 µm-filtered U conc.). The BCF in stomach (w/content) was 18.5 L/mg (particulate U conc.). The other organs did not reach steady state concentrations within 28 days of exposure.

The accumulated concentrations in internal organs (muscle, kidney and liver) documented U being transported in the blood. Uptake of U through skin and stomach were not excluded in aqueous exposure of U to Atlantic salmon. Results demonstrate that even longer exposure are needed to identify steady state.

Key Words: Uranium, Atlantic Salmon (Salmo salar), Uptake-rate, U-accumulation, U- concentration, U-fractions, bio concentration factor

(17)

Sammendrag

Masteroppgaven omhandler økotoksikologiske effekter av uran i økosystemer i ferskvann, ved å analysere de mulige opptaksveiene av uran i fisk ved å benytte ung (3 mnd gml.) Atlantisk laks (Salmo salar) som modellorganisme. Biokonsentrasjonsfaktor (BCF), opptakshastighet og konsentrasjonsfordeling i organer ble benyttet for å bestemme det potensielle opptaket av uran.

Hypoteser: (1) Uran blir tatt opp i fisk direkte fra vann. (2) Konsentrasjonen av uran i gjeller er høyere enn i mage ved likevekt ved eksponering i vann. (3) Opptakshastigheten i gjeller og skinn er større, sammenlignet med muskel, nyre og lever.

Atlantisk laks (n=50) ble eksponerer for vannbåret uran i moderat hardt US-EPA vann i 28 dager. En stabil pH (6.7) og stabile betingelser for vannet medførte at U spesieringen var konstant. Vannprøver ble samlet hver dag og fraksjonering av vann ble utført tre ganger underveis i eksponeringen (dag 2, 17 og 28). Det ble tatt prøver av fiskene fem ganger i løpet av eksponeringen. Selve prøvetakingen fulgte EMERGE protokollen (Rosseland et al., 2001).

Bioakkumuleringen av U ble målt de forskjellige vevene. Konsentrasjoner av målte grunnstoff ble kvantifisert ved bruk av Agilent Technologies 8800- QQQ ICP-MS.

Presisjonen og nøyaktigheten ble bestemt basert på analysen av de sertifiserte

referansematerialene, de estimerte relative standard avvikene (RSD), og kvantifikasjons- og deteksjonsgrense.

Den generelle vannkvaliteten var ikke signifikant forskjellig mellom kontrollen og 50 µg/l uran-eksponeringen. Den gjennomsnittlige konsentrasjonen av uran i den nominelle 50 µg/l- eksponeringen var 55 ± 22 µg U/L (n=32) og 0.027 ± 0.038 µg U/L (n=38) i kontrollen. Den lavmolekylære fraksjonen av uran tilsvarte 19 % av den totale konsentrasjonen. Den antatte biotilgjengelige (U kation) konsentrasjonen var i gjennomsnitt 1.13 ± 1.74 µg U/L av den totale uran-konsentrasjonen i vannet. Tilsetning av U, jevnlige bytter av eksponeringsvann og rask binding av biotilgjengelig uran til løst organisk materiale (DOM) skapte en dynamisk forandring i fraksjoneringen av U.

Etter 28 dager var akkumuleringen av uran signifikant i alle de analyserte organene. Det var ikke et signifikant opptak av U i kontrollene. Den største U konsentrasjonen og

(18)

gjelle/dag). Konsentrasjonene av uran var i synkende rekkefølge: gjeller, mage m/innhold, skinn, nyre, muskel og lever. Opptakshastighetene var signifikant høyere for de organene med direkte kontakt med det kontaminerte vannet, sammenlignet med de indre organene som ble analysert (muskel, nyre og lever). Den tilsynelatende BCF i lever var 0.52 L/mg (<0.45 µm- filtrerte U kons.). BCF i mage m/innhold var 18.5 L/mg (partikulær kons.). De andre organene oppnådde ikke likevekt i løpet av de 28 dagene med U-eksponering.

De akkumulerte konsentrasjonene i de indre organene (muskel, nyre og lever) dokumenterte at uran ble transportert i blodet. Opptak via skinn eller mage ble ikke ekskludert i Atlantisk laks gjennom vanneksponering av U. Resultatene viste at lengre eksponeringstid er nødvendig for å kunne oppnå likevekt.

Nøkkelord: Uran, Atlantisk laks (Salmo salar), opptakshastighet, uran-akkumulering urankonsentrasjon, U-fraksjoner, biokonsentrasjonsfaktor

(19)

1.Introduction 1.1 Uranium

1.1.1 Radioactivity

Natural uranium consists of three isotopes ²³⁸U, ²³⁵U and ²³⁴U (Barillet et al., 2007). There is 99.2745 % of ²³⁸U, 0.7200 % of ²³⁵U and 0.0055 % ²³⁴U (Choppin et al., 2002a). The half-life of ²³⁸U is about the same as the age of the Earth (4.5 x 10⁹ years), which states that this isotope is not very radioactive (Bleise et al., 2003). The specific activity of ²³⁸U is 12.4 MBq/kg and 25.4 MBq/kg of natural uranium (Choppin et al., 2002a). ²³⁸U is the parent isotope of several more radiotoxic elements like thorium (Th), plutonium (Pu), radium (Ra) and radon (Rn). All of the daughter nuclides have shorter half-life than the ²³⁸U isotope. The final product in the U decay chain is lead (²⁰⁶Pb), which is stable (Bourdon et al., 2003).

235U (0.72 % of natural uranium) is fissile, i.e. capable to undergo fission, which is the

production of electricity occurring in a nuclear reactor. Before U may be used as nuclear fuel, it has to undergo an enrichment process where the ²³⁵U percentage is increased to about 3 %.

Before the enrichment process starts, U is required to be in gaseous form and is then

converted to a fluoride. After the process is finished, the uranium fluoride (UF6) is separated to enriched U and low enriched U, also known as depleted uranium (Choppin et al., 2002b)

1.1.2 Sources of U

Several sources have contributed to concentrations of radioactivity in the environment and naturally occurring radioactive materials (NORMs). Weathering of U- and Th- containing minerals is a natural event and nuclear weapon and fuel cycle (U mining), and oil and gas- industries are all anthropogenic or man-made activities (Salbu et al., 2015). The

anthropogenic nuclear weapon and fuel cycles are the main sources of U contamination to the environment (Goulet et al., 2012). Production and testing of nuclear weapons generate a lot of radioactive waste. Atmospheric testing of nuclear weapons is the main source of plutonium in the environment, which is known to be a highly radioactive element (Finch et al., 1999).

The nuclear fuel cycle is divided into three steps, the front-end, the nuclear power station and the back-end part. The front-end includes mining, milling and enrichment, and the back-end includes radioactive waste treatments (Choppin et al., 2002). Mining, milling and refining

(20)

generates the greatest volumes of radioactive waste. The spent nuclear fuel is the largest source of radioactivity (Finch et al., 1999).

1.1.3 Depleted uranium

U and DU have the same behaviour in the body, but DU is about 60 % less radioactive (Bleise et al., 2003). The concentration of ²³⁵U in depleted uranium is 0.2-0.3 %, compared to about 0.7 % in natural uranium (UNEP, 2001). Due to the low radioactivity, DU is commonly used to analyse mechanisms and model uptake of natural uranium (Song et al., 2014).

DU is a biproduct of the pre-treatments of natural uranium (Lind et al., 2020). In several countries, DU is used in military purposes as armour plating and ammunition (Bleise et al., 2003). To characterize U as DU, the concentrations of ²³⁵U and ²³⁴U are reduced relative to the concentration of ²³⁸U. The weight of ²³⁵U has to be 0.2 % of the total mass of DU, and the weight of ²³⁸U has to be 99.8 % (WHO, 2001). The atom-ratio between the isotopes

(²³⁵U/²³⁸U) in DU is reported to be 0.002 (Lind et al., 2020).

1.2. Uranium in water

1.2.1 Naturally concentrations of U

The concentration of U in water is dependent on the geological material surrounding the water source (WHO, 2001). In 2005, aqueous concentrations (lakes, rivers) of total uranium were determined in several countries (Australia, Canada, Kyrgyzstan and Kazakhstan) with large U reservoirs, to detect possible U contaminations from the mining activities. Natural

background U concentration is quantified as <1 µg/L. In Australia all concentrations were below natural background (n= 525). About 75 % (n= 68 303) of the quantified U

concentrations in Canada were below natural background concentration. The remaining 25 % ranged up to 1350 µg U/L, as a result the clean-up technology was improved, and the

concentrations got closer to natural background in 2009. In central Asia (Kyrgyzstan and Kazakhstan), the U concentrations in the rivers reached maximums of 3.1 µg/L (n= 14) and 41 µg/L (n= 160) (Goulet et al., 2012). Pit lakes (U mining sites) in central Asia (Kyrgyzstan, Kazakhstan and Tajikistan) had U concentrations ranging from 7.8 µg/L to 3 mg/L (Lind et al., 2013; Salbu et al., 2013; Skipperud et al., 2013b; Strømman et al., 2013).

In 1996, the Norwegian Institute of Water Research (NIVA) conducted a study determining the concentrations of different elements in Norwegian lakes (n=475). The concentration of U

(21)

was in the range of 0.004-2.22 µg/l (Skjelkvåle et al., 1996). Several Norwegian

groundwaters (n=426) were analysed to detect the concentration of different elements (Be, Th, U, Cd, Hg etc.) in 2000. The water was sampled from boreholes in Norwegian crystalline bedrocks. The concentration of U in 18 % of the samples exceeded the American maximum admissible concentration of 20 µg/l. The median was 2.5 µg U/l and the highest quantified concentration was at 750 µg U/l (Frengstad et al., 2000).

1.2.2 Speciation of U in water

The natural low concentration of uranium in rivers and lakes originates mainly from the erosion of rocks and minerals (Komperød et al., 2015). In water, the metal exists as U(IV) or U(VI), and the form is dependent on the redox potential in the water. The U(IV)-species are almost insoluble and normally found in sediments, while U(VI)- species are in general mobile and soluble (Goulet et al., 2012). In a reducing environment, U(VI) is reduced to U(IV), likely due to microbial involvement (Windom et al., 2000).

1.2.3 Factors affecting the speciation of U in water

The bioavailability of a metal and the further toxicity in natural water is highly dependent on the physio-chemical form of the metal (speciation) (Franklin et al., 2000). The speciation of U in water depends on the pH, dissolved organic material and dissolved phosphorus (Goulet et al., 2012).

The speciation of U is directly linked to pH. The uranyl-ion (UO22+) may form complexes with either hydroxy-(OH-) or carbonates- group(s) (CO32-) when the pH >5, in oxic freshwater (Goulet et al., 2012). At pH 6.7, the primary low molecular mass U species are:

UO22+,UO2CO30, UO2(OH)20 and UO2OH⁺, as well as other complexes with hydroxy- or carbonate-ligands (figure 1.1). pH 6.7 is commonly observed in Norwegian lakes (Solheim et al., 2018).

Dissolved organic material (DOM), or dissolved organic carbon (DOC), are strong ligands, which may bind to dissolved uranium in freshwater (Trenfield et al. 2011). DOM has a

negative charged surface, which make cations in the water easily sorb to the surface and cause an in-direct inhibition of accumulation onto other surfaces in the same environment (mucus, fish gills etc.) (Rosseland, 2000). Higher concentrations of DOM and DOC may decrease the concentration of the uranyl-ions (UO22+), and increase the concentrations of colloidal- and

(22)

particulate U. DOM forms stable complexes with the uranyl-ion in freshwater (figure 1.1, Goulet et al., 2012), and particulate U in freshwater sediments is affected by deposition of DOM (McManus et al., 2006; Chappaz et al., 2010). The complexation of organic matter and metals increases with pH and the solubility of metals decreases with pH, which both affect the formation of metal-ligands in freshwater with a pH closer to neutral (Franklin et al., 2000).

Figure 1.1: The graphs indicate the fraction of total UO2 bound to DOC at different concentration affected by pH (Goulet et al., 2012).

Phosphorus (P), or phosphate, at high concentrations in surface waters has the possibility to form complexes with U and decrease the fraction of bioavailable U species. Uranium may form complexes with phosphate ions if the concentrations are >0.1 µg/l, which is not assumed to occur in natural freshwater (Goulet et al., 2012).

The distribution of U species changes depending on available ligands. Uranium may form complexes with sulphate or fluoride (>1 % of present species) if the pH is low (<6.2). The uranyl ion is often referred to as the “free U ion”, but it is in fact coordinated by water molecules (Goulet et al., 2012).

(23)

Figure 1.2: Speciation of LMM-U linked to pH. Uranium concentration in A) is 1nmol/L an B) is 110 nmol/L (Goulet et al., 2012).

1.2.4 Fractionation in water

The physio-chemical properties define the radionuclide species, for instance oxidation state, charge properties and valence, density, structure, degree of complexation and nominal

molecular mass (Salbu, 2007). The U compounds can be divided into different fractions. The definitions by size are that particles have a diameter larger than 0.45µm, colloids or

pseudocolloids have a diameter 1 nm-0.45 µm, and LMM-species or low molecular mass species have a diameter less than 1 nm (Salbu, 2007). Colloidal masses are larger than 10kDa and LMM species are less than 10kDa (Teien, pers comm). The U in pit lakes in central Asia were mainly present as LMM-species (DellaSala and Goldstein, 2017; Skipperud et al., 2013b)

In a similar study analysing U chemotoxicity was the predominantly part LMM-species, and about 30 % U colloids (Teien et al., 2014). The LMM- species are believed to be the most mobile and bioavailable fraction, i.e. can be transported by active uptake (across cell

membranes). Particles and colloids are considered to be inert biologically (figure 1.3) (Salbu, 2007). The colloids are considered to be mobile in water. The particles are believed to be the most immobile fraction, which precipitates in water and is likely found in the sediments (DellaSala and Goldstein, 2017). All the fractions consists of different U species (Goulet et al., 2012).

(24)

Figure 1.3: Overview of fractions (Salbu et al., 2015)

1.3 Uptake of uranium in fish

1.3.1 Bioavailability of U

In aqueous exposure, U(VI) is more favourable than U(IV) to be taken up in the fish.

However, U(IV) may sorb to the particles in the sediments and be eaten by fish (Goulet et al., 2012). The uranyl-ion has a positive charge, which makes it likely to bind to the negatively charged mucus and surface of gills like other cations (ex.: Al³⁺) (Rosseland, 2000). Other LMM-U is less bioavailable due to a less positive, neutral or negative charge (figure 1.2) (Goulet et al., 2012).

The pH may affect the accumulation of U in two ways: by changing U speciation or the H⁺- concentration. Franklin et al. (2000) studied the toxicity of U in algae at pH 5.7 and 6.5. The results revealed different inhibition of growth in the two exposures. pH 6.5 was identified as the most toxic pH, and because the U speciation were fairly similar in the two exposures, the U toxicity was linked to the H⁺-concentration where the H⁺-ion inhibited the uptake of U at the lower pH (Franklin et al., 2000).

The uranyl-ion (UO22+) is considered to be the most bioavailable specie of U (Goulet et al., 2012). It has been assumed that high concentrations of phosphate in water reduces the bioavailability of U. However, the phosphate concentration is predicted to be well above 60 µg/L in order to make an effect of the toxicity of U (Goulet et al., 2012) .

(25)

1.3.1. Uranium toxicity

Natural and depleted uranium are considered to have low radiotoxicity (Giblin et al., 2015). U has a larger potential as chemotoxic than radiotoxic (Mathews et al. 2009), because of the low specific activity (Simon and Garnier-Laplace, 2005)

The activity of the free hydrated metal-ion is considered to affect the metal toxicity in water.

The total metal concentration is less relevant to estimate the toxicity (Franklin et al., 2000).

There is little information about the U toxicity to fish, but U is considered to be one of the less toxic metals. The extent of chronic effects of U exposure have limited information, and most of the mechanisms are probably unknown (Goulet et al., 2012).

1.3.2 Organ distribution

Bioaccumulation of U in fish is documented in previous studies (Barillet et al., 2007;

Lerebours et al., 2009; Song et al., 2014; Simon et al., 2019). The metal does not biomagnify in the food chain (Goulet et al., 2012). Within fish tissue, U generally accumulates in

mineralized tissues like bone and scales. Smaller concentrations may be detected in kidney and liver. Detectable concentrations are likely to be determined in the gill filaments, skin and muscle. Concentrations of U have been detected in the brain, which makes U one of few metals which pass the blood-brain barrier (Barillet et al., 2007; Goulet et al., 2012; Song et al., 2014).

The World Health Organization (WHO) states that U(VI)-compounds may become adsorbed through the skin in humans (WHO, 2001). There is little information about U sorption to fish skin, or skin as a possible uptake-route. In Barillet´s study (2007), the highest concentrations of U were detected in the liver and gills, when the same organs as listed above were analysed.

U accumulates heterogeneously in fish, which is likely explained by the organs different physiological roles in uptake and transport of U (Barillet et al., 2007; Lerebours et al., 2009).

U is often bound to bicarbonate when transported in the blood. The kidney filters the blood, and this may cause higher concentrations of U in this tissue (Goulet et al., 2012). The kidney is considered as the main target organ in humans (WHO, 2001). The U accumulation depends further on the concentration of bioavailable U, typically reactive LMM U species, and the total U concentration is not directly related to the accumulation (Song et al., 2014).

(26)

A mass balance approach was conducted to demonstrate in which tissues U generally

accumulates. This model predicts about 20 % of the U uptake to accumulate in the bones, 18

% in remaining hard tissues, 2.8 % in gonads, 0.6 % in kidney, 1.3 % in liver, 43 % in muscle and 15 % in remaining soft tissues (Yankovich, 2009). This model has not been validated in the field (Goulet et al., 2012)

In the same pit lakes as previously listed (Kazakhstan, Tajikistan and Kyrgyzstan), U

concentration in fish were studied. The results showed U accumulation in gills, liver, kidney, (bone)¹ and muscle. The U distribution varied between the fish species, but generally the highest accumulation was quantified in the gills. At similar U concentrations in the water in Kazakhstan and Tajikistan, the highest accumulations of U were in the liver and gills. The pH at the two sites were 8.5 and 8, respectively (table A.1, Appendix 1) (Salbu et al., 2013;

Strømman et al., 2013; Skipperud et al., 2013b; Lind et al., 2013).

1.3.2.1 Uptake through the gills

Few studies have analysed the biokinetics of uptake of uranium through the gills (Goulet et al., 2012; Lerebours et al., 2009; Simon et al., 2019). Ions like calcium and magnesium have the same chemical potential as U. It is likely that uranium can be taken up through the same transport mechanisms as the active uptake of Ca and Mg in the gill membrane (Lerebours et al., 2009; Giblin et al., 2015).

Lerebours et al. (2009) completed an experiment with U uptake in zebrafish. The total concentrations of U in the water, as well as in organs, were quantified. Concentrations in the gills and liver were not significantly different throughout the exposure (4000 ng U/g dw). The concentration in skeletal muscles was quantified to be about 700 ng U/g dw (Goulet et al., 2012; Lerebours et al., 2009). The steady-state of U uptake was not reached in this experiment (Lerebours et al., 2009). The uptake appeared to increase asymptotically over the 28 days of exposure, by following a single-compartment first order kinetics (Goulet et al., 2012).

Uranium accumulates in fish gills, and the accumulation correlates positively with the

concentration of U in the water. The uptake is dependent on the pH; in the range of 5.5-6.0 is the accumulation to the gill high and in the range 7.3-7.7 is the accumulation lower (Giblin et al., 2015). The uptake through the gills may decrease if the uranyl-ion is blocked by the

1 Bone was not included in the analysed tissues in this study.

(27)

protective mucus-layer on the gills surface. This blocking-effect has been revealed for other divalent metals (Barillet et al., 2007; Lerebours et al., 2009).

Lead (Pb) has similar traits as U (Nieboer and Richardson 1980). A study conducted by Hodson (1978), compared uptake routes and accumulation of lead in different organs for rainbow trout (S. Gairdneri). The main uptake was from water and through the gills, and the central accumulation was in the bones, gills and kidney. Uptake from dietary Pb-exposure was low compared to the aqueous exposure, and was not above control levels (Hodson et al., 1978).

1.3.2.2 Uptake through the gut

Few reports describe how U is taken up through the gut in fish from contaminated food. U is a reactive trace element, which may sorb to the surface of particles and colloids. The fish can eat these objects, which may cause a detectable U concentration in the stomach. Some data indicate that fish living close to sediments (benthic) have an uptake of U through the stomach (Goulet et al., 2012).

Bleise et al. (2003) referred to previous studies conducted by Harley et al. (1999) analysing the uptake of U in animals. The uptake from the ingested DU contaminated food was about 2- 5 % from the intestines to the blood. The rest of the U in the food was assumed to just pass the intestines and not be taken up in the body. The major concentration (90 %) of U in the blood was predicted to be excreted though the kidneys, within a week. The last 10 % was assumed to accumulate in the organs, whereabout 15 % was predicted to accumulate in the bones (Bleise et al., 2003). The same trends for uptake through the gut in animals may be seen in fish too, but this is not known. In humans were the uptake-percentage of U from

contaminated food about 2 % (WHO, 2001).

Simon et al. (2019) conducted a similar study analysing the uptake-routes in waterborne (20 µg/L U) and dietary (10.7 µg/L, ²³³U (radioisotope) exposure of U on zebrafish (D. rerio).

The aqueous exposure was either performed alone or in combination with the dietary. The timespan was 5 or 20 days. The pH was 6.5 ± 0.7. The total U concentration in the water was in the range of 21-22 µg/L. The intestines had the highest U accumulation after 20 days of waterborne exposure compared to the other organs analysed (gill epithelium, liver, kidneys,

(28)

gonads, and intestine). The concentrations in the intestines (10-15 µg/g tissue fw) were about thirty times higher than the quantified concentrations in the gills after 20 days.

The U concentration provided by the water was two times higher than the concentration provided by the diet. Simon et al. included the aspect of absorption of waterborne U to the uncontaminated food, but the food was consumed by the fish fast (1-5s) which reduced the contact time. The present decrease of U in both exposures were not significantly different.

The measured diet-borne transfer of the radioactive U isotope (²³³U) was quantified as low, which reduced the weight of intestine contamination via food consumption (Simon et al., 2019).

In another study, Crayfish (O. Limosus) were fed with U contaminated bivalves. The trophic transfer order differed between individuals (1-13 %). The U accumulated in the stomach and digestive gland and reached concentrations of about 12 µg/g fw and 18 µg/g fw respectively (Simon and Garnier-Laplace, 2005).

1.3.2.3 BCF of uranium in fish

The bioconcentration factor is the transfer from water to fish (Lind et al., 2013). The BCF may only be calculated when the uptake of U reaches a steady state, i.e. the uptake

concentration in the organ is the same as the concentration eliminated. The uptake of U over time will likely slow down (Bleise et al., 2003).

𝟏𝟏. 𝐵𝐵𝐵𝐵𝐵𝐵= 𝑈𝑈 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 (𝑓𝑓𝑐𝑐𝑓𝑓ℎ) µ𝑔𝑔 𝑘𝑘𝑔𝑔 𝑑𝑑𝑑𝑑

𝑈𝑈 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 (𝑑𝑑𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐)µ𝑔𝑔/𝐿𝐿=𝐿𝐿/𝑘𝑘𝑔𝑔 (Skipperud et al., 2013a)

There are few calculated BCFs describing the uptake of U in fish. According to Goulet et al., it is quantified values from 0.001 to 0.149 L/kg. The concentrations stem from several studies (Barillet et al., 2007 etc.) with different water chemistry and various fish species. Another uncertainty is the fact that U is usually measured in some tissues and not the whole body- concentration (Goulet et al., 2012).

The bioconcentration factors were quantified in tissues (liver, gill, muscle and kidney) in Kazakhstan, Tajikistan and Kyrgyzstan (table 1.1). The BCF varied between tissues, but

(29)

generally the highest BCF was quantified in the gills. At similar U concentrations in the water in Kazakhstan and Tajikistan, the highest BCFs were quantified in the liver and gills (Salbu et al., 2013; Strømman et al., 2013; Lind et al., 2013).

Table 1.1: Overview of quantified BCFs at Kazakhstan, Tajikistan and Kyrgyzstan.

Site U conc. in

water

pH TOC

(mg/L)

BCF Reference

Kazakhstan 1.3 mg/L 8.5 1.76 Liver: 2.5 L/kg (dw) (Salbu et al.,

2013;

Strømman et al., 2013).

Gill: 2.9 L/kg (dw) Muscle: 0.11 L/kg (dw)

Tajikistan 1.4 mg/L 8 2.23 Liver: 5.6 L/kg (ww) (Strømman et

al., 2013) Gill: 3.6 L/kg (ww)

Muscle: 0.12 L/kg (ww) Kidney: 5.2 L/kg (ww)

Kyrgyzstan 41 µg/L 7.6-

8.7

1.48 Liver: 0.25-1.6 L/kg (ww) (Lind et al.

2013) Gill: 0.63-1.9 L/kg (ww)

Muscle: 0.043-0.10 L/kg (ww)

1.4 Biotic factors affecting the uptake of U in fish

1.4.1 Biodilution

A potential growth during exposure may cause a biodilution of the U concentration in the fish.

The biodilution factor ought to be low to be able to reach a steady state in U uptake (Teien, pers comm).

1.5 Atlantic salmon (Salmo salar)

The average fertile salmon (about one year old) weighs between 1-3 kg. The salmon is an anadromous specie, which has the ability to live in both rivers and in the ocean. The fish spawns in freshwater and lives most of its life in the ocean (Hansen, 2000). Atlantic salmon (S. Salar) is one of the most sensitive fish species, which makes the salmon a favoured specie to use in an ecotoxic study (Poléo et al., 1997). The mucus works as an immune system for the fish, by covering external surfaces when the concentration of a pollutant reaches toxic levels for the fish (Barillet et al., 2007; Rosseland, 2000; Teien, pers comm) . Rapid changes in the ecosystem, for instance spring floods, may be critical to the juvenile salmon and cause fish death (Rosseland, 2000).

(30)

Atlantic salmon is a central fish species in Norway, with both cultural and economic aspects, which makes it important to maintain the fish health by determining possible pollutants in studies. Norway is the world's leading producer of salmon, with more than half of the world's production ("Laksefakta", Seafood Norway).

1.6 Objectives

When identifying the uptake of U in fish, the uptake-rate and the bioconcentration factor (BCF) are important factors to determine. These factors may be used to develop models to assess the ecotoxic effects of uranium in freshwater ecosystems.

Three hypotheses were tested in this study:

(1) There is an uptake of uranium in the fish directly from water;

(2) The concentrations of uranium in gills is higher than in stomach at steady state in aqueous exposure;

(3) Uptake-rates in gills and skin, compared to muscle, kidney and liver are higher;

The main goal of the experiment was to identify the uptake-rates of U in fish using juvenile Atlantic salmon (Salmo salar) as a model organism.

(31)

2.Method

2.1 Fish holding conditions

This master thesis was a part of a Ph.D. study, which compared uptake and depuration of uranium from waterborne U and U contaminated food. The MSc. part of the experiment focused on how fast U was taken up from waterborne exposure and the following distribution of U between tissues. Due to the extent of the experiment, only one treatment (50 µg U/L aqueous exposure) was evaluated and compared to the control in this thesis. The entire experiment followed OECD guideline 305 for bioaccumulation in fish (OECD/OCDE, 2012) for a period of 28 days. The experiment was approved by the Norwegian Animal Research Authority (FOTS ID: 19370)

2.1.1 Atlantic salmon

Atlantic salmon (Salmo salar) juveniles from the fish laboratory of Norwegian University of Life Science (NMBU) were used in the experiment. The eggs were obtained from Aquagen AS (Trondheim, Norway). The salmon were fed about 1-month post-hatch (before exposure).

The feed was a yeasted-based pellet that met the required nutrients for the fish. This feed had low concentrations of U (0.055 ± 0.001 µg/g feed). The fish were maintained in RAS

(recirculating aquaculture system) lab water during early development life-stages.

One week prior the exposure start, 50 fish (3-month-old, average weight: 1.2 g, average length: 4.8 cm) were transferred from a batch holding tank to exposure vessels and kept in US-EPA moderately hard water for acclimatization before the start of the exposure.

Moderately hard water was produced from deionized water in batch-tanks (800 L) using standard recipe (table 3.3) and aerated for 24 hours prior to use.

The US-EPA moderate hard water was modified to reach a pH at 6.7 instead of 7.4-7.8 to ensure a larger fraction of the uranyl-ions to be present in the water (Goulet et al., 2012, figure 1.2). The concentration of NaHCO3 (sodium bicarbonate) which originally was set for moderate hard water was replaced with the concentration set for soft water. The NaHCO3 was the main salt affecting the pH, because the increased concentration of carbonate ions (CO32-) in the water increased the pH. The concentration of uranyl-ions may have been affected by the excess carbonate-ions, due to increased sorption. Sodium chloride (NaCl) was added to reach the listed concentration of sodium in the water, which increased the concentration of chlorine

(32)

(Cl) in the water. The rest of the salts added, followed the set standards for US-EPA moderate hard water.

All of the water used during acclimatization and during the experiment was adjusted to pH 6.7 and kept at 15 ⁰C. U exposure water was made using preconditioned moderate hard water, which was transferred to a separate tank (220L). U was added from a stock as uranyl hexahydrate (UO2(NO3)² x 6 H2O) (Sigma Aldrich) to a final concentration of 50 µg U/L.

The chronic low concentration of U in this experiment provided a constant exposure, which was not assumed to cause toxic conditions and further affect the uptake of U in fish. The U contaminated water was stored 48 hours before being transferred to the fish tank, to ensure stable water quality.

The fish were fed twice a day at 2 % bodyweight during acclimatization and throughout the entire course of experiment. The fish needed to be fed during the 28 days of exposure, as exposure without feed can only continue for some few days.

2.2 Exposure system

2.2.1 Design

Exposure tanks comprised a recirculating flow-through design and each exposure (control and 50 µg U/L) was performed in duplicates. Atlantic salmon juveniles (n=50) were placed in duplicate 25 L-tanks for both exposures (control and 50 µg U/L). Tanks containing fish were connected to a header tank (4 x silicon tubes, 5 mm inner diameter), which provided a continuous water flow (figure 2.1).

Overflow from the fish tank entered the 100 L tank, via a CO2-stripper, to be recirculated.

The CO2-stripper was conducted by PVC (polyvinyl chloride) tubes filled with high surface- area plastic inserts. Freshly aerated water was provided to the header tank by a submersible water pump from a 100 L tank below the fish tank. The design of this system provided constant waterflow at a stable water level to the fish with oxygen saturated water with low CO2. Each tank had a lid to keep the evaporation-rate as low as possible.

(33)

Figure 2.1: Fish (n=50) were transferred from the breeding-tank to a separate exposure- tank. The tanks were connected by tubes to maintain a continuous waterflow and a stable level of water in the exposure tank.

2.2.2 Water quality

It was important to minimize possible stress factors other than U. One goal was to ensure optimal water quality for the fish during the experiment. Temperature, pH, conductivity, CO2- level, dissolved oxygen, NH3/NH4+-concentration were then measured throughout the course of the experiment. The pH, conductivity and O2-saturation were measured using WTW- Multi 340i and - Multi 3420. A climate-controlled room maintained constant temperature at 15⁰C.

The CO2-level was measured by Oxyguard CO2 analyser and the NH4-level was measured by a Merck spectrophotometer.

Introduction of feed into the system increased the concentrations of ammonium (NH4+) by excretion from fish. Ammonium can be transformed to ammonia (NH3) in water, which may be toxic to fish at high concentrations. To ensure low concentrations of NH3, the

concentration of NH4 was quantified regularly. It was assumed that the feed contributed to increase the concentrations of DOM in the water, which might further increase sorption of U to DOM colloids and particles. The exposure water was changed about two times per week to decrease the concentration of DOC. All water was removed from the tanks (except the fish tank). Tanks were scrubbed clean and fresh water (either control or 50µg/L U) was

reintroduced within 20 mins of drainage to not stop the circulation of water in the fish tank for too long.

CO2-stripper

Water pump

O2-pump

(34)

The pH was adjusted to 6.7 by adding HCl (4 M) in the synthetic EPA-water before use. The pH was not adjusted after the new water was transferred to the fish tanks, but daily

measurements detected any change from the nominal value. Oxygen-saturation was kept close to 100 %. An O2-pump was installed to keep this concentration of oxygen stable. The

concentration of LMM U was maintained during the exposure by spiking the water with a U stock (UO2(NO3)² x 6 H2O) (Sigma Aldrich). The U concentration had 40 % renewal every day (20µg/L), i.e. daily U-spikes.

2.3 Sampling

2.3.1 Water-samples

Water samples were collected throughout the experiment, to determine day-to-day variation in U concentration and fractionation. From early on, a rapid loss of bioavailable U in the

exposure water was determined.

In addition to pre-exposed U measurements, the concentration of U was quantified almost daily throughout the experiment. The filtration methods used for water samples were 0.45µm- filtration, ultrafiltration and chelex. The total concentration of U in the tank was quantified through analysis of the unfiltered samples. The <0.45 µm-filtered samples did not include particles and thus separated the uranium bound in particles from the water. U colloids and LMM U species were then present in the filtered fraction. Unfiltered and filtered samples were collected every time the water was changed in the tanks. The unfiltered samples were collected with a pipette, and the filtered samples with a 0.45 μm-filter and a 10 mL syringe.

Uranium fractionization using ultrafiltration and ion chromatography were in addition used at day 2, 17 and 28 to separate colloids, low molecular mass species and low molecular mass- ions. Unfiltered, filtered and samples for the <10kDa and <10kDa chelex were collected and paired samples were completed. The pre-treatments for the <10kDa and <10kDa chelex were performed straight after water sampling. An ultrafiltration (UF) was completed of the

<10kDa-samples. The unfiltered water samples were then run through a tube filled with hollow fibres. The ions and low molecular mass uranium (LMM U) weigh less than 10kDa and were then able to move through the ultrafilter in the hollow fibre and be collected in the filtrate. The LMM U species consisted of ions and low molecular mass complexes.

(35)

To determine the content of reactive ions of the ultrafiltration, paired water samples were run on the chelex. The chelex was a tube filled with a resin (chelex 100 resin, Bio-Rad). The resin contained particles, which had sodium-ions bound to the surface. The sample was led through the tube, and the more electron-negative ions than sodium switched position with the Na⁺-ion.

All positive charged U species in the sample was assumed to accumulate in the resin. A larger quantity of sodium in the <10kDa-chelex than in the <10kDa confirmed ion levels in the sample. If uranium was quantified in the sample after the chelex-treatment, complexes of uranium were present.

The samples were stored in 15mL-tubes. The water samples were stored in a temperature of +4^oC, to decrease the possibility of evaporation and fouling.

Calculations of the different U fractions:

Unfiltered sample = Total uranium

Unfiltered – 0.45 µm filtered = particulate uranium 0.45 µm filtered – <10kDa = colloidal uranium

<10kDa = LMM U species

<10kDa – <10kDa chelex = U cations

2.3.2 Fish samples

To determine the uptake and distribution of U in juvenile salmon, fish from each duplicate exposure were dissected on day 0, 2, 4, 8, 17 and 28. At sample point, fish (n=3) were transferred to tank (1L) containing Finquel anesthesia (MS-222 (Tricaine mesylate),

100mg/L, Scan Aqua AS), before measuring weight and length (figure 2.2). The length was measured from nose-tip to the end of fishtails. A picture was taken of each fish when the length was measured.

Blood samples were collected to determine glucose level (mmol/L). Changes of glucose-level in the blood indicated that the fish suffered from stress. The skin, gills, liver, kidney, stomach (w/content), brain, muscle and bone were dissected following the EMERGE protocol

(Rosseland et al., 2001). When sampling the fish from the aqueous exposure, it was important to not contaminate the organs by contact with the skin. The equipment (metal tweezers of two different size, scalpel and a scissor) was cleaned in between tissues by using a paper towel.

Aluminium foil was used to maintain a clean workspace, this foil was changed in between different exposures. The equipment was cleaned more thorough between exposure groups

(36)

with EtOH (70 % vv) to avoid cross-contamination. The samples were placed in pre-labelled 5 mL tubes and stored at -20⁰C.

Figure 2.2: The left picture, length was quantified of fish. Right picture shows the fish before the dissection of internal organs (Pictures: B. Terum, 11.06.19).

2.4 Tissue and water analyses

2.4.1 Pre-treatments: water

All the water samples were sorted by date and given a unique number. Every volume was adjusted to 13 mL by using a pipette. The unfiltered and <0.45 µm-filtered samples did not undergo any pre-treatments in the lab.

Ultrapure HNO3-acid was added to each sample to reach 10 % concentration (1.1 mL). The reference material was 1640a. Ten blank tests were pre-made. A house standard was used in the analysis (1643H). The water samples were further run on ICP-MS, together with the standards and the CRM.

2.4.2 Pre-treatments: fish

The frozen samples where freeze-dried overnight. The dried samples were weighed on a scale, by placing the tissue on the lid of the tube. The risk of contamination was then minimized. The sample was moved with either a plastic tweezer (organ) or a plastic pipette (only stomach). The equipment was rinsed with EtOH (70 %) between each organ. New equipment was used for every sample-point. The samples were weighted in increasing concentrations, to minimize the risk of contamination from the equipment.

(37)

2.4.2.1 Digestion

100 µL internal standard and 500 µL of ultrapure HNO3-acid was added to each sample prior to the digestion. The samples were further placed in an oven at 90^oC for an hour to digest the tissues. After the digestion, the samples were diluted with 4.5 mL distilled water, i.e. filtered reverse osmosis (RO)-water. The final volume was 5 mL. The pre-treatments were then completed.

2.5 Data analysis

2.5.1 ICP-MS

The analysis was run on an Agilent ICP-MS QQQ – 8900 (mass spectrometry). The ICP-MS used ≈2 % HNO3 as rinse solution between the samples and a 5 % HNO3-solution as a liquid carrier. Either O2-gas or He-gas was used as carrier-gas.

The results from the measurements on the ICP-MS were corrected based on analysis of the internal standard, online standard and quantifications of one known solution (drift). The internal standard detected loss of analyte (U) in the pre-treatments. The online standard was analysed regularly throughout the analysis, to correct for effects of matrix and drift in the instrument. The drift is a change in the instrument’s measurement during analysis series, which cause a lower determined concentration in the sample.

The certified reference materials (ERM BB-422, IAEA 414 and 1640a) were used to analyse the accuracy of the measurements. A house standard (1643H) was analysed together with the water samples. See appendix 2 for further details about the analysis. Data explaining the concentrations of total organic carbon (TOC), chlorine (Cl) and nitrate (NO3-) during the exposure were not ready when this thesis was delivered. The weights were confirmed by plotting rubidium and total mass of the organs.

3.5.2 Formatting data and statistics

The results from the analysis on the ICP-MS were transferred to Excel spreadsheets. The drift was detected by analysing a known solution regularly throughout the analysis and calculated as a factor². This factor was multiplied with the quantified concentrations in the unknown

(38)

samples. The detection limit (LOD) and quantification limit (LOQ) were calculated. Precision of the method was quantified by calculating relative standard deviation (RSD) of the

measurements.

The most optimal gas-carrier was chosen for each element. Then the results of the controls and 50 µg/L were isolated. The averages and standard deviations of both exposures were calculated for each day. The final results were plotted. To quantify the level of accuracy of the method, deviation of an CRM (certified reference material) and a house standard (1643H) were determined. Precision of the method was quantified by calculating relative standard deviation (RSD) of the measurements.

The exponential equations describing the uptake-rates in the different organs were calculated using SigmaPlot. The uptake-rate was assumed to follow a first order kinetic function with exponential rise to the maximum. The accumulation was assumed to have a rapid increase during the first days of exposure (approximately linear trend) and further flatten out, which indicated the concentration of uptake and elimination approached steady state. In this study, the uptake was assumed to reach steady state within 28 days of exposure.

A is the concentration after x days. A0 represents the steady state concentration. The K-value is the uptake coefficient. The uptake is affected by the already accumulated concentration of U in the tissue at the time. The uptake-rate is assumed to be large in the beginning of

exposure and then slowly decrease over time until reaching the steady state concentration (Teien, pers comm)

2. A = A0 (1- e^-kx)

(39)

3. Results/discussion

3.1 Quality of the analysis on the ICP-MS

The concentrations of U in all the organs and water samples were above the quantification limits. The average LOQ³ for organs was 0.00026 µg/g based on average 0.01g sample. LOQ was 0.0011 µg/l for water samples. If the concentration in the tissue was predicted to be low (example: muscle), a larger part of the organ was sampled (if possible) to lower the

quantification limit on the ICP-MS. The total concentration of the element in the sample was then >LOQ.

3.1.1 Quality of the ICP-MS analysis (water)

The quantified concentrations and the certified concentrations of elements in the certified reference material (1640a) and the house standard (1643H) were listed in table 3.1. The results indicated good precision as the variations in the quantified concentrations of the different elements were low (RSD < 5 %), except for sulphur, in the house standard. The bias (%) were in the range of the nominal values (1643H) or <3 %, i.e. good accuracy. The

determined concentrations of elements in certified reference material had a bias <3.2 %, also acceptable.

Table 3.1: Overview of the reference material used when measuring the water samples with certified concentrations and determined concentrations of house standard and CRM. Values presented in average ± std.dev. Calculations of RSD and bias in appendix 2 (eq. 7 and 4).

Ions 1643H 1640a

Nominal value

Quantified

value ⁴ RSD%

Bias

%⁵

Certified value

Quantified

value ⁶ RSD%

Bias

% Na

(mg/L) 20.7±0.26 19.9±0.3 1.4 -1.2 3.14±0.03 3.01±0.04 1.2

-3.2 Mg

(mg/L) 8.0±0.10 8.08±0.09 1.1

Within

1.059±0.004 1.08±0.07 6.5

2.9 P (mg/L) 2.5 2.42±0.02 0.8 -2.4

S (mg/L) 2.5 2.5±0.2 6.9 Within

K (mg/L) 2.03±0.029 2.009±0.008 0.4 Within 0.580±0.002 0.558±0.001 1.1 -3.0 Ca

(mg/L) 32±1.1 31.2±0.1 0.5

Within

5.62±0.02 5.4±0.3 6.2

-3.1 U (µg/L) 1 1.02±0.03 2.7 Within 25.35±0.27 25.0±0.6 2.2 -0.8

3 Calculations of LOD and LOQ in Appendix 2 (eq. 8 and 9)

4 N=4 for 1643H

5 “Within” means within the range of house standard.

(40)

3.1.2 Quality of the ICP-MS analysis (fish)

Two CRMs were used to determine the accuracy of the U concentration in fish samples. The bias was <5 % for all of the certified elements analysed.

Table 3.2: Certified concentrations and determined concentrations of the reference material, in addition

deviation (bias, %) in certified reference material used when analysing fish samples. Values presented in average

± std.dev. Calculations of RSD and bias in appendix 2 (eq. 7 and 4).

Ions ERM BB- 422 (g/kg)

ERM BB- 422-mea.⁷

RSD (%)

Bias (%)

IAEA 414

IAEA 414- mea.⁸

RSD (%)

Bias (%)

Na (mg/L) 2.80 2.81± 0.03 1.1 0.36

K (mg/L) 21.4 20.5± 0.6 3.0 - 4.21

Ca (mg/L) 0.342 0.33± 0.02 5.2 -2.94

U-238 (ng/g)

86 - 92.5 77.5 ± 6.1 7.9 - 2.8

3.1.3 Discussion of quality of the ICP-MS (water and fish)

All the CRMs, both for water and fish, showed a bias (%) <5 %, which was an acceptable level of accuracy. The variation in quantified concentrations were low <5 % (RSD). The low values of RSD indicated good precision in the analysis of water and fish. Determination of uranium measurement (IAEA-414) showed RSD at 7.9 %, which indicated a larger variation in IAEA-414 standard than in the other CRMs. The dry-matter content in the IAEA-414 was 94.4 %. It would be preferred to use at least two CRMs for the determination of uranium to confirm the quantified U concentrations, especially if the CRM had a larger RSD than 5 %. In this experiment, it was important with a correct quantification of the U concentration in the samples. By comparing two CRMs, which both were certified for U, the credibility of the overall analysis would likely increase.

3.2 Characteristics of aqueous exposure

This chapter presents the water characteristics of the exposure water and the stability of the water chemistry throughout the experiment.

7 N=5 for ERM bb-422

8 N=6 for IAEA-414

Uptake of uranium in Atlantic salmon (Salmo salar) : aqueous exposure

Uptake of uranium in Atlantic

salmon (Salmo salar) – aqueous

exposure

Uptake of uranium in Atlantic salmon (Salmo salar) – aqueous exposure

Acknowledgements

Statement of declaration

List of acronyms

Table of Contents

List of tables

List of figures

List of appendices

Abstract

Sammendrag

1.Introduction 1.1 Uranium

1.2. Uranium in water

1.3 Uptake of uranium in fish

1.4 Biotic factors affecting the uptake of U in fish

1.5 Atlantic salmon (Salmo salar)

1.6 Objectives

2.Method

2.1 Fish holding conditions

2.2 Exposure system

2.3 Sampling

2.4 Tissue and water analyses

2.5 Data analysis

3. Results/discussion

3.1 Quality of the analysis on the ICP-MS

3.2 Characteristics of aqueous exposure